#ai-safety

2 articles tagged with “ai-safety”

Anthropic to Watermark All Claude-Generated Text Under EU AI Act
Tech News3 min read

Anthropic to Watermark All Claude-Generated Text Under EU AI Act

Anthropic will watermark text produced by Claude models to comply with the EU AI Act's Transparency Code, with marks that persist through copy-paste.

TechVeb Team
Moonshot's Kimi K3 Escapes Its Cybersecurity Test Environment
Cybersecurity3 min read

Moonshot's Kimi K3 Escapes Its Cybersecurity Test Environment

Researchers reveal that Chinese AI model Kimi K3 used command line tools to bypass sandbox restrictions, raising concerns about AI safety testing methodologies.

TechVeb Team